GitHubEmploy

#DOCUMENTAION

Deep Learning Coronavirus Cure

The goal is to create a novel small molecule which can bind with the coronavirus, using deep learning techniques for molecule generation and PyRx to evaluate binding affinities.

Binding scores of leading existing drugs (HIV inhibitors) are around -10 to -11 (the more negative the score the better), and around -13 for the drug Remdesivir which recently entered clinical testing.

By combining a generative RNN model with techniques and principles from transfer learning and genetic algorithms, Was able to create several small molecule candidates which achieved binding scores approaching -18.

Acknowledgements

GitHubEmploy

class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(1, 64, kernel_size=(3, 3), padding=1)
        self.conv2 = nn.Conv2d(64, 64, kernel_size=(3, 3), padding=1)
        self.max_pool = nn.MaxPool2d(2, 2)
        self.global_pool = nn.AvgPool2d(7)
        self.fc1 = nn.Linear(64, 64)
        self.fc2 = nn.Linear(64, 10)

GitHubEmploy

import torch
import torchvision
import torchvision.transforms as transforms

num_epochs = 10
batch_size = 32
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

GitHubEmploy

# MNIST dataset
train_dataset = torchvision.datasets.MNIST(root='data',
                                           train=True, 
                                           transform=transforms.ToTensor(),
                                           download=True)

test_dataset = torchvision.datasets.MNIST(root='data',
                                          train=False,
                                          transform=transforms.ToTensor())

GitHubEmploy

optimizer = torch.optim.Adam(model.parameters(), lr=0.001)
loss_function = nn.CrossEntropyLoss()

GitHubEmploy

model.to(device)

GitHubEmploy

# Train the model
total_step = len(train_loader)
for epoch in range(num_epochs):
    for i, (images, labels) in enumerate(train_loader):
        # Move tensors to the configured device
        images = images.to(device)
        labels = labels.to(device)

        # Forward pass
        outputs = model(images)

GitHubEmploy

with torch.no_grad():
    correct = 0
    total = 0
    for images, labels in test_loader:
        images = images.to(device)
        labels = labels.to(device)
        outputs = model(images)
        _, predicted = torch.max(outputs.data, 1)
        total += labels.size(0)
        correct += (predicted == labels).sum().item()

GitHubEmploy

torch.save(model.state_dict(), 'model.ckpt')

GitHubEmploy

// ==UserScript==
// @name		Geoguessr Cheat
// @namespace		https://www.leonbrandt.com
// @version		2.0.0
// @description		Press SHIFT + ALT + G and the location of your current geoguessr game will open in a new tab
// @author		Leon Brandt
// @homepage		https://www.leonbrandt.com
// @updateURL		https://gist.githubusercontent.com/leonbrandt/16b3a70ef70939359357c908e6b0f06d/raw/geoguessr-cheat.user.js
// @match		http*://*/*
// @grant		none